From: DougAtHome (dcalton@fuse.net)
Date: Fri Feb 21 2003 - 16:37:57 GMT-3
I have been playing around with DLSw+ in my lab, and using netbios traffic to
practice with. Without going into specifics, the general setup for DLSw might
look like this on my routers:
dlsw local-peer peer-id 170.10.5.1
dlsw remote-peer 0 tcp 170.10.3.1
dlsw bridge-group 1
interface Ethernet0/1
ip address 170.10.12.1 255.255.255.192
bridge-group 1
bridge 1 protocol ieee
For LAN switching, I am using a 2950 with fairly recent code, and have set up
vlans for the router ethernet ports and the PCs generating the netbios
traffic, via the "switchport access vlan #" command.
If I do nothing else (and no examples I have ever seen suggest that I need
anything else), the 2905 ports automatically switch to trunk mode (apparently,
the default administrative mode is "dynamic desireable", and of course
everything goes to Hell in a handbasket. I can correct this by configuring
the switch's port to access mode (nonegotiate), but this got me to wondering
what was really going on with bridging.
How exactly does bridge protocol ieee end up convincing the switch that it is
talking to a trunk link and what would be a better way to resolve this issue?
My first idea was to include "bridge-group 1 spanning-disabled" on the
router's ethernet interfaces, but this didn't help. I also briefly played
around with going full bore and configuring the switch ports as trunks with
appropriate native vlans, but of course this would have led me down the path
of IRB, which is clearly incorrect. Anyone have any thoughts or theories, or
can better explain what is going on here? TIA!
This archive was generated by hypermail 2.1.4 : Sat Mar 01 2003 - 11:06:28 GMT-3